Samulski 1990

INT. I. HYPERTHERMIA, 1990, VOL. 6, NO.
5, 909-922
Clinical experience with a multielement ultrasonic hyperthermia system:

analysis of treatment temperatures
T. V. SAMULSKI, W. J. GRANT, J. R. OLESON, K. A. LEOPOLD,
M. W. DEWHIRST, P. VALLAFUO and J. BLIVIN
Duke University Medical Center, Radiation Oncology, Box 3085, Durham, NC 27710,
USA
(Received I1 July 1989; revised 20 December 1989; accepted 3 January 1990)
Int J Hyperthermia Downloaded from informahealthcare.com by McMaster University on 02/03/15
A summary of turnour temperature data obtained from 3 1 patients who underwent 147
hyperthennia treatments with the Sonotherm lo00 ultrasonic system is presented. The
treatment goal was to achieve a minimum of 42-0°C in turnour for 60 min duration
with normal tissues remaining below 43.0"C. In 83% of treatments at least one measured
turnour temperature reached or exceeded 42.OoCat some time during the treatment.
Nineteen per cent of these treatments had a time- and spatial-averaged temperature
(measured in tumour) 242.O"C. A variety of anatomical sites were treated and these
were grouped into four categories: groin/trunk,axilla, breadchest wall and head/neck.
Measured temperatures in turnours located in the groin and trunk sites were significantly
higher (22% 242°C) than other locations. The head and neck treatment temperatures
were significantly lower (8% of measured pointsr42"C).
For personal use only.
1. Introduction
The lhitations of inducing hyperthermia non-invasively using electromagnetic (EM)
fields are well appreciated (Hand and James 1986). A severe limitation arises from the
dependence of the EM field attenuation on the wavelength in tissue. For wavelengths smaller
than the typical target dimensions (i.e. 1 cm in high water content tissue), the field
-
attenuation is very high ( 6 dB/cm). Therefore, localized heating can be accomplished
at shallow depth and only very superficial tumours can be effectively heated. For deep-
seated targets, relatively longer EM wavelengthsare needed to increasethe field penetration.
This necessarily results in regional energy deposition since the wavelengths are much larger
than the typical target volume.
Despite the above fundamental limitation most hyperthermia therapy has been delivered
using EM techniques and most hyperthermia equipment available from commercial sources
are EM devices. Ultrasound (US) has better potential for locally heating target tissues
at intermediate(3-5 cm) and deep-seated (5-15 cm) sites. This potential again arises from
the field attenuation versus wavelength relationship. The US wavelengths useful for
therapeutic heating are of the order of millimetres or less, and are much smaller than the
typical target dimension. Additionally, the attenuation in tissue (- 1 dB/cm/MHz (Hand
and James 1986)) allows better plane-wave energy deposition. This combination of small
wavelength and good field penetration allows geometric focusing that can produce localized
power deposition at depth.
There have been reports of the use of US in therapeutic hyperthermia (Marmor er al.
1979, Corry er al. 1982, 1984, 1988, Kapp et al. 1988, Shimm et al. 1988, Lele 1983,
1986, Coleman er al. 1986). The equipment used in these clinical studies was, by and
large, fabricated at the respective institution. The progression towards the commercial
development of US equipment for hyperthermia has only recently occurred. There are
now several products that have recently completed the pre-market approval phase of clinical
trials to demonstrate that explicit US devices are safe and effective.
02566736/90$3.00 01990 Taylor & Francis Ltd.

910 i? V. SQmUlski et al.
This work reports the clinical experience with a commercial US hyperthermia system
(Labthermics Sonotherm loo0 System, Labthermics Technologies, Inc., Champaign, IL,
USA) at Duke University. The report is limited to the characterization of hyperthermia
treatments with this system on a population of 31 patients by analysing the measured
temperatures obtained in tumour. In order to quantify the hyperthermia achieved in these
clinical applications, an analysis in accordance with that reported by the equipment
evaluation study groups sponsored by NCI contract is used ( C o w er al. 1988, Kapp et
al. 1988). The results of therapeutic response and treatment toxicity will be presented
in a separate work.
2. Materials and methods

2.1. Sonorhenn loo0 system
The Sonotherm loo0 system is a product of Labthermics, Inc. Technical aspects of

a prototype system are described by Underwood er al. (1987). The radiofrequency (r.f.)
hardware consists of 16 independent amplifiers (channels) capable of delivering 130 W
of electrical power in each channel. Transducers of two sizes are available with the system.
These transducers are fabricated from single, square, PZT-8 ceramic piezoelectric crystals.
The larger transducer has an effective radiating area of 15 cmx 15 cm, and the smaller
transducer has an effectivearea of 7 cm X 7 cm. These transducers are further subdivided
into smaller square elements or segmentsof size 3 * 5 cm X 3 * 5 cm (Underwood et al. 1987).
Thus, the larger applicator is a transducer with 16 segments and the small applicator has
four segments. These small segments are individually electrically matched to one of the
driving amplifiers and therefore can be powered separately by the respective amplifier.
The design goal is to allow individual power control to the various segments while
maintaining minimum cross-coupling among the segments and minimum variation in the
near-field power intensity distribution (Underwood er al. 1987). This segmented transducer
design allows some degree of geometric field shaping through the choice of which segment
to power-activate. It also allows power intensity modulation among the various activated
segments that can be useful for achieving better temperature uniformity, as well as
minimizing treatment limiting pain symptoms.
Two US frequencies can be used with these applicators: the fundamental frequency
- -
mode at 1 MHZ and the harmonic at 3 MHz. This choice in frequency adds a dimension
of depth control over the US energy deposition profile. All segments are operated at the
same frequency at any given time. The system does not allow for automated modulation
between the two operating frequencies. However, current control software does allow for
the manual change of frequency during treatment.
The system has a temperature feedback control system with software control and display
implemented on a PDP 11/35 computer (Digital Equipment Corp., Merrimack, NH, USA).
A detailed description of the control algorithm is not published. In practice one specifies
the location of surface and interstitial thermocouple sensors within the treatment field using
a surface-projected, two-dimensional alphanumeric grid. The grid is geometridy
correlated with the segments of the US transducer. The sensors are identified with respect
to the tissue type (normal or tumour). Limits are assigned which delineate the temperature
range desired for tumour tissue (e.g. 42°C minimum and 45°C maximum) and the
maximum allowed temperature for normal tissue (e.g. 43°C). The tissue categorization,
respective temperature limits and grid location co-ordinates are input parameters required
at the time of treatment set-up. The temperature control algorithm modulates power to
the gwmetrically appropriate transducer segments in order to keep the tumour temperature
sensors in the specified therapeutic range and normal tissues below the maximum toIerance
limit.
Multi-element ultrasonic hyperthennia system 911
2.2. lhemmetry
The Sonotherm lo00 system includes a 16-channel copper-constantan thermocouple
thermometry system. The resolution of temperature measurements is better than 0.1 "C
and accuracy of this same order can be routinely achieved after suitable system warm-up
-
( 30 min) using a single fixed-temperaturepoint calibration (figure 1). The thermocouple
probes used in these treatments were suppliedby a separate source (Sensortek, Inc., Clifton,
NJ, USA). Stainless-steel copper-constantan thermocouple needles (23 gauge) with one,
two or three sensors were used without catheters for invasive temperature measurements.
Flexible vinyl-covered, single-sensor (0.64mm OD) probes were used for surface
measurements. Prior to each treatment the 16 channels of the system were calibrated in
a water bath referenced to an NBS traceable mercury-in-glass thermometer using the surface
thermocouple probes. The thermocouple probes and the thermometry channels have proved
to be interchangeable within an accuracy of *0-2"C (see figure 1). Thus the invasive
needle probes were sterilized and used without recalibration at the time of treatment. If
a probe or sensor indicated signs of error or malfunction prior to or during a treatment,
this sensor was rechecked for performance integrity post-treatment. If a probe or channel
proved to be in error the data are excluded from treatment analysis.
A second thermometry unit (Luxtron UXK)fibreoptic system-luxtron, Mountain View,
CA, USA) has been interfaced with a previously described automatic temperature Scanning
(ATS) system (Engler et al. 1987). Single-sensor fibreoptic probes (model MPM)were
used in conjunction with 20-gauge, blind-end, stainless-steelneedles (lengths 9 or 15 cm)
to allow linear temperaturetracking. This ATS data acquisition system was used in addition
to the thermocouple system in the latter one-half of the treatments in this study.
The combination of the glass fibreoptic probe and stainless-steel needle was used to
reduce artefact associatedwith absorptionand viscous heating. The steel needie encasement
prevents selectiveUS absorption that would occur with the bare fibreopticprobe or a plastic
14 -
12 -
10 -
(r
0
8-
6-
4-
k
2-
O+
-0.l<x50.0 O.O<X50.1 O.l<XSO.2
Difference in Temperature ("C)
-
Figure 1. A comparison of thermocouple probe readings after single point calibration at 42 .O"C
of the Sonotherm lo00 thermometry system using a mercury-in-glass standard thermometer
and water bath. The frequency distributions of differences between probe readings (Taand
thermometer standard (T,)are plotted. Twenty-seven probes am randomly chosen and
distributed among the system channels after the single-point calibration procedure.
912 T. V. Samulski et al.
mapping catheter. The viscous heating artefact was characterized for this fibrwptic probe
and steel needle combination in a fresh beef phantom. The artefact was estimated by
extrapolating the linear rise in temperature with time back to the time point when US power
was supplied to the phantom. The displacement in temperature, AT, is the assumed extent
of artefact and has been reported as a value normalized by the rate of temperature change
of the bulk medium (i.e. "C/"C per min) (figure 2a). Additionally, the fibre/needle probe
combination was tested for thermal conduction smearing in a step temperature gradient
(figure 2b). The experimental technique for assessing thermal conduction error is described
by Lyons et al. (1985). No corrections for either viscous heating artefact or thermal
smearing have been made in the data prior to analysis.
Before the implementation of the thermal mapping data acquisition system an average
of seven invasive thermocouple sensors were used per treatment. After the mapping system
was introduced an average of 21 measured intratumoral spatial temperature points per
treatment were monitored. The average time between automated maps was 5 min. The
distance between points along the linear maps was 0 . 5 cm. In addition to the automatic
mapping data an average of three fixed thermocouple sensor points per treatment was
recorded using the Sonotherm lo00 system. Thermometry data collected with both the
fixed thermocouple sensors and the thermal mapping systems have been used in the analysis
of results. Surface temperatures recorded with flexible thermocouple probes are not included
in the data analysis.
2.3. Patient population

Patients having biopsy-proven primary, recurrent or metastatic malignancies of various

histologies and anatomical sites were eligible for enrolment in clinical protocols which
use hyperthermia as an adjuvant to radiation or chemotherapy. Patients were selected for
hyperthermia treatment by US provided the target lesions were considered appropriate.
The criteria for selection included lesions that: (1) were accessible for the external placement
of the US applicator, (2) could be adequately encompassedby the expected energy deposition
pattern in terms of area ( 515 c m x 15 em) and depth ( 58 cm), (3) had acoustic windows
that were largely devoid of bone and intervening air spaces, (4) were accessible for invasive
temperature monitoring, and (5) were considered to be not better treated with alternative
heating devices (e.g. microwave applicator or interstitial implants). As a result of this
last criterion the majority of lesions in this study were generally large with respect to
diameters and depth [average volume (product of maximum orthogonal diameters x d 6 )
-
=230 cm', average depth =5 8 cm] .
Thirty-four patients were enrolled for treatment with the Sonotherm lo00 system. Thirty-
one patients had treatment durations and temperature measurements available for analysis.
The 31 patients analysed underwent 147 treatments. All patients signed an informedconsent
statement approved by the Duke University Institutional Review Board. A summary of
patient data is included in table 1.
The treatment objectives were to elevate tumour temperatures to a minimum of 42-0°C
for a 60 min duration while maintaining normal tissues below 43°C. Therapeutic time
was started when any measured intratumoral point reached 42 -0°C. If no measured tumour
temperature reached 42-0°C within 15 min after power was initiated, the start of therapeutic
time was taken to be 10 min after power-on. Power was increased as quickly as possible
until the patient experienced persistent pain that was not tolerable, or until normal tissues
reached 43°C.
A majority of patients were given pre-treatment medication in order to reduce anxiety
or partially compensate for treatment related pain. These medications included morphine
sulphate (8-10 mg), brazepan (1-2 mg), meperidine hydrochloride (50-75 mg) and
diazepam (5-10 mg). No patient, however, received treatment if not alert.
Multi-element ultrasonic hyperthermia system 913
6-
5-
F
G
a
4-
2
5
E 3-
i
Y
2-
r
I-
!
f 1-
0 1 MHz
01 S = 4A'/min.
I I I 1
-1
o 10 20 30 40 50 60 70 80 90
T
E
M
P
E
R
A
T
U
R
E
37
0 1 2 3 4 5
POSITION (cm)
@)
Figure 2 . Test for viscous heating and conduction smearing with the MPM fibreoptic probe inside
a blind-end, 2O-gauge, stainless-steel needle. An example of viscous heating artefact is shown
in (a). The sharp rise AT=0-4ZoC at the initiation of US power is indicative of artefact.
This artefact for the fibre/needle combination is characterized by AT/S where S is the rate
of temperature rise for the bulk medium (i.e. AT/S=O.W"C/"C per min in this example).
The measurement of a step-upstepdown tempemhue distribution is shown @) for the fibreoptic
probel2O-gauge needle combination and the fibreoptic probe/l9-gauge, thin-wall Teflon cathekr
combination (seeLyons et al. 1985 and samulski et al. 1985). This test demonstrates conduction
errors for the fibdneedle combination are comparable to the fibdcatheter combination
routinely used in M W hyperthermia thermometry.
Table 1. Sitelsize of tumour for patients in study
’ Histology Sizelvol. (cm3) T,
-
T,,, (“(3
Groinhunk
1 Malignant 8 X 8 X51168 7 41 -73
8 Transitional cell carcinoma 5 x 5 x4152 3 39.57
14 Liposarcoma 13 x 13 x 81707 9 41 *40
18 Squamous cell carcinoma 8x 3 ~ 5 1 6 3 3 40.58
22 Fibrosarcoma 10x 7 x 71257 2 38-58
24 Thyroid carcinoma 15X 16 X 9/113 1 5 42.54
27 Renal cell carcinoma 4 x 4 x5142 2 37-47
29 Fibrous histiocytoma 7 x 14x 101513 5 39.11
30 Synovial cell sarcoma 1Ox5X71183 10 42-53
Axill&
3 Schwannoma 8 x 7 x 71205 6 40.20
6 Malignant melanoma 5 X 8 X5/105 4 42-32
17 Adenocarcinoma 6 x 7 x 61132 4 40.02
19 Fibrous histiocytoma 8 x5 ~71147 8 40.08
20 Adenocarcinoma 4 X 5 X 4/42 5 40.72
Breadchest wall
10 Malignant melanoma 3X2X319 5 40.72
12 Adenocarcinoma 7X11X51202 7 41 -25
15 Adenocarcinoma 4 X 3 X 2/13 2 37 * 74
+diffuse nodules
16 Adenocarcinoma 10X 10X 10/524 9 40.31
23 Adenocarcinoma 6~6x3157 2 40.90

26 Adenocarcinoma 7 x 13 ~ 8 1 3 8 1 10 40.79
28 Malignant melanoma 14x 8 x 81469 3 -
39 62
Headheck
2 Squamous cell carcinoma 8 x 4 x 71117 1 39-89
4 Malignant melanoma 13 X 12 X 5 / 4 8 2 37-81
5 Leiomysarcoma 5~6x4163 5 39.77
7 Squamous cell carcinoma 15 x 8 ~61377 2 40.21
9 Squamous cell carcinoma 6x5~3147 1 38.15
11 Squamous cell carcinoma 6x6~61113 4 41-58
13 Squamous cell carcinoma 15 x 8 x 61377 5 39-56
21 Squamous cell carcinoma 4 X 4 X2/ 17 3 39-8
25 Squamous cell carcinoma 6 x 5 x 6/94 3 38.81
31 Squamous cell carcinoma 6x5~71110 10 39.76
‘=Chronological-order of tient accession.
v o ~=volume
. (a*b*c)*zX’
-
T,=Total number of treatments included in analysis.
T,,,=Mean of T,,, over all treatments.
2.4. Treatment set-ups

Because of the enhanced flexibiiity with regard to the multi-segmentedtransducer and
feedback control, considerable attention to treatment set-up is required with the Sonotherm
system. Several treatment set-up techniques were used and modified with growing clinical
experience. The descriptionsof immobilizationand alignment techniqueshave been reported
elsewhere (Oleson and Samulski 1989).
Coupling the US transducer to the tissue surface is a critical component of the treatment
set-up. Several methods of coupling the US applicator to tissue were used. These include
the taut elastic membrane, closed-volume conformingor ‘floppy’membrane and the open-
volume conforming or ‘floppy’ membrane (figure 3). The taut elastic membrane was the
least useful since it is only appropriate for coupling into concave tissue contours. The
Multi-element ultrasonic hyperthennia system 915
a
elastic
membrane
b closed volume
“floppy’
membrane
3
‘coupling
gel
open volume
‘I1oppy‘
membrane
\coupling ’
gel
Figure 3. Techniques used for coupling the US transducer to the tissue surface: an elastic convex
membrane (a), closed-volume ‘floppy’membrane (b) and open-volume ‘floppy’membrane (c).
Each technique has advantages and disadvantages. The latter two techniques were used for
the majority of patients in this study.
closed-volume ‘floppy’ membrane (figure 3b) and open-volume floppy membrane (figure
3c) approaches were much better with regard to coupling the irregularly shaped body
contours. The majority of treatments were carried out using these latter two coupling
configurations.In all cases ultrasound coupling gel was applied between skin and coupling
membrane.
Another important set-up requirement is the documentation of invasive probe
placements. For each patient’s initial treatment set-up, invasive probes were placed using
computed tomographic (CT) guidance and verification to localize the probe with respect
to tumour and normal tissue. Robe positions were reproduced as closely as possible on
subsequent treatments. When changes in tumour geometry (due to therapy response or
oedema), or proximity of the probe to critical structures occurred, CT guidance and
verification were used as necessary. The localization of probe positions relative to the
external transducer was accomplished using a transparent marking template as described
by Oleson and Samulski (1989).
3. Data analysis
Thermometry data acquired from this patient group have been analysed in terms of
the percentage of a temperature-related parameter greater than or equal to a given index
temperature (T&) (Kapp et al. 1988). The temperature parameters include average
916 T. V. Samukki et al.
treatment temperature, average maximum and average minimum treatment temperature.
The average treatment temperature, T,,, is defined as the average of all intratumoral
temperatures measured during the 'therapeutic treatment time'. The parameters T- and
T,, are the spatial point in tumour with the time-averaged maximum and time-averaged
minimum temperature, respectively, achieved during treatment. In some instances
treatments were stopped prior to the prescribed 60 min therapeutic time owing to patient
tolerance or equipment malfunction. For these cases averages were calculated from the
start of therapy (i.e. first point reached 42°C or 10 min after power-on) to the time of
power-off.
The percentage of temperatures (967)2 Tndex was - determined
-- for all measured
intratumoral points. Grand averages (tables 1 and 2-T,,,, T-, Tmin) are individual
temperature parameters averaged over all treatments.
Patients were grouped according to anatomical site (table 1): headheck, axilla,
breastfchest wall, and groinlbody trunk. The temperature parameters were compared for
these respective site groupings. Differences in temperature parameters based on these group
classifications were tested for significance using the t-test.
4. Results
4.1. Thennometry errors
Viscous heating artefact was estimated for the fibreoptic and stainless-steel needle
combination. The average artefact was found to be AT=0-2"loC per min (range
0.01-0-35°C) and AT=O-O3"C/"C per min (range 0.0-0.l"C) at 1 MHz and 3 MHz,
respectively. The artefactwas estimated by extrapolatingthe linear temperaturerise response

back to the time when US power was initiated and normalized by the rate of temperature
rise for the bulk medium (see figure 2a). Typical rates of temperature rise achieved during
treatment are 0-5-2 -O"C/min. Thus, one can estimate average viscous artefacts to be on
the order of 0.1-0-4"C. There is no mechanism within the Sonotherm software to measure
or correct for artefact arising from the fixed thermocoupleprobes. Consequently, artefacts
for the fixed-needle probes have not been characterized in this work. Similar probes have
been studied elsewhere (Fessenden ef al. 1984) and the artefact is expected to be of the
order of 0-2°C during these treatments. Since artefacts for both the fixed and mapping
probes are difficult to detect, and highly dependent on many factors (e.g. power incident
at the probe site, probe orientation relative to incident beam, the probe-tissue interface)
no correction for this possible artefact has been made in the data.
The results of the step gradient thermal conduction smearing test are shown in figure
2b. The combination of Luxtron MPM fibreoptic probe tracking through the stainless-
steel needle and the same fibreoptic probe tracking through a 19-gauge, thin-wall Teflon
catheter are compared. The curves are identical within the error of the experimental
procedure. In a previous work (Samulski et al. 1985) it was shown that a fibreoptic probe
tracking within a small gauge Teflon catheter yielded less thermal smearing than other
probehtheter combinations. Thus the thermal conduction error using this MPM
probe/20-gauge, stainless-steel needle combination is comparable to that associated with
fibreoptic/19-gaugeTeflon catheter combination and no corrections for thermal smearing
are made to the temperature data.
4.2. Temperature analysis

In 83% of the treatments at least one intratumoral temperature probe reached 42.0"C
during the treatment time. The average number of measured points per treatment that
reached 42 -0°C was 7 -6out of an average number of 17 3 measured points. The average
time these points were at or greater than 42-0°C was 24 min.
Table 2. Grand averages f standard deviation for various thermal parameters and treatment subgroups.
5i$3
- - - No. of points Ave. time 2
No. of treatments TaveoC TmoC Td,"C No. of points Tx time (min) >42"C >42°C 9
m
Whole Group 147 40.44 1*92 42.73 *2-31 38*50* 1-28
* 17-30* 12.33 54.99 12.42 7.62 f 9 *18 23.94* 17.70
* z
Large appl. 97 40-54*2- 15 43-06*2.43 38.47*1.33 19.58f 13.67 55.46* 11.79 9.59* 10.33 26.81 f 17-00 E
Small appl. 50 40-25 1* 33 42.08 1-87 38 -54f 1 17
* - 12.88 *7 * 12 54.06 f 13-52 3 * 82 *4.37 18-38f 17-69 83
Less 15Occ 65 39*96*2*07 41*98* 1.88 38.5Ort 1.24 11.98*6.98 53.98* 14-12 3.69*4.36 19*98*18.35 i;.
More 15Occ 82 40 * 83 1-69 43.32 *2*44 38.49 f 1 *31
* 21 -51f 13-93 55 -78f 10.83 10.74 f 10.68 27-09f 16.50
preATS; no map 34 40 * 11f 2 - 5 9 42- 15f 1-86 38.81 f 1-29 6.91 * 5 . 0 4 50.84 f 15.72 3.18 f3.65 20.71 f 17-24 Q
preATS 18 40-34zt2-32 42-59zt1.97 38.72f1.18 10.63zt7.49 54.19f13.75 4.44f4.66 25*51*18-70 &
ATS 95 40.50 1* 65 42-80*2 -47 38 -37f 1-31
* 20.95 12.93 55 -42z t 11*60 9.37 f 10.49
* 23 -08 17-06
*
Groinltrunk 46 41 -08* 1*82 43*68&2*67 38.59* 1-52 21 -35f 15.65 55*46* 11* 18 11*67* 12.28 28*85* 17-02 3
a.
Axilla 27 40.55 f 1* 10 42.5 1 f 1-72 38 -70i 1*04 14.78f 8.85 58 19f 8- 84 4.32 i5 -29
- 23 -55 19.41
*
Breast1C.W. 38 40.26zt2.44 42.81 *1.96 38.37*1.09 19*79*11*68 58*26*7.21 9*24*7.63 26-60*14-90 3
Headlneck 36 39.73 f 1.59 41*58* 1.95 38-35* 1-25 11*39*6*20 48.53 f 17.20 3.22*4-26 15.28 16.70
* 3
The percentages greater than or equal to TdCxfor all measured intratumour points
made during the 'therapeutic time', T,,, T, and Tmin, are shown in figures 4 a,b,c,d,
respectively. The proportion of points142"C was 18%. The proportions of treatments
with an average, maximum and minimum temperature142°C were 19%,58 % and 1% ,
respectively.
Figure 5 shows the distributions for %Tz Tadexwhen separated by treatment site. The
average temperature achieved in the headneck treatment site is found to be significantly
(ps0.05) lower than the other treatment sites when compared separately. The average
temperature achieved in the groin/trunk site was significantly higher (when compared
separately) than those in the other treatment sites @ ~ 0 - 0 0 1 )The
. difference between
average temperature achieved in the headneck and groin/trunk sites was found to be
significant @ ~ 0 . 0 5when
) compared by total number of treatments as well as when
compared by number of patients within these site categories.

A summary of results for grand average thermal parameters for - the -various
classifications of treatment groupings is shown in table 2. Differences in T,,, T,, and
are not significant except when analysed by anatomical site.
Temperature measurement techniques evolved over the course of this study. Initially
temperatures were measured using fixed thermocouple probes only (table 2 -preATS; no
map). Later treatments had both fixed probes and manual thermal mapping (table
2-preATS). Subsequently the combination of fixed and automatic mapping techniques
-
imn
L %TzTinda l o r d patknts
31 p.(ienta 147 tmtmentr
m.0 - L
I m.0-
=
F
A
z
a0 -
0.0
Figure 4. Distributions of measured tumour temperatures plotted as %T> Tkx for all patients
(31 patients, 147 treatments). The distribution of all tumour temperatures measured during
the 'therapeutic time' for a l l treatments is shown in (a). The distribution of average treatment
temperatures T,, is shown in (b). The distribution of average maximum (minimum)
temperature per treatment T,, (Tmi,,)is shown in (c) and (a), respectively.
Multi-element ultrasonic hyperthema system 919
Figure 5. The %T> T- for hunour temperatures measured during the therapeutic time for
treatments grouped by anatomical sites. (a) Groin/trunk (nine patients, 46 treatments); (b)
breastkhest wall (seven patients, 38 treatments); (c) axilla (five patients, 27 treatments); (d)
h d n e c k (10 patients, 36 treatments). The distributionfor the gmin/trunk site is significantly
better (e.g. higher %T>T* at 42 and 43°C)than other sites. The distribution for the
h d n e c k site is significantly worse than other sites.
were used (table 2-ATS). The average number of spatial points (table 2-No. of points)
monitored per treatment using these various temperature sampling techniques probes was
seven (preATS; no map), 11 (preATS) and 21 (ATS). There was no difference in the
calculated thermal parameters based on these sampling variations.
Differences in calculated thermal parameters based on tumour size stratification
(<150 cm3 versus> 150 cm3 were not significant (table 2). Finally, there was no
difference in thermal parameters based on whether the large or small applicator was used
(table 2).
5. Discussion
In 83%I of treatments given with the Sonotherm lo00 system at least one point in tumour
reached 42~0°C at some time during the treatment session. On average, 7 - 6 measured
points per treatment achieved this temperature level and were maintained at 42-0°Cor
higher for an average of 24 min (table 2). This 42:O"C tumour temperature threshold
was found to be a clinically significant prognostic treatment criterion by Sapozink et al.
(1986) in their examination of hyperthermia delivered by an EM regional device.
As is generally the case for most IocaVregional clinical hyperthemia applications,
regardless of the heating technique, the ability to achieve the temperature prescribed in
the treatment protocol (i.e. 42°C minimum tumour temperature for 60 min) is limited.
This was also our experience with the Sonotherm lo00 US system. The proportion of
treatments with a space-time-average temperature2 42 .O"C was 19% in this patient group.
This benchmark value is lower than those reported by other investigators using the US
modality (Kapp et al. 1988,Cony et al. 1988). There are differences between institutional
reports with regard to thermometry data acquisition and analysis which can account for
this apparent discrepancy. In this study a relatively large number of temperature points
-
per treatment (average no. 17) were monitored and, as shown by Corry et al. (1988),
there is a general trend towards lower average temperature parameters with an increase
in the number of temperature points monitored. Despite this fact it appears that the
percentage of points reaching 42.0-43.0"C is low in this data base.
Complaints of pain during treatment were common, and these pain symptoms were
often treatment-limiting. It has been reported elsewhere that pain limiting symptoms are
more prevalent using the plane-wave US applicator than using the unfocused microwave
applicators (Kapp et al. 1988). This experience appears to support that observation. It
is important to understand that the differences in depth of penetration for plane-wave US
and superficial MW devices are significant, and the consequenceof increased pain symptoms
using unfocused US may be the result of heating significantly larger normal tissue volumes.
There are also common complaints of referred pain, particularly when treating the axilla.
In some cases the flexibility offered by the segmented applicators allowed the identification
of areas which were symptomatic for pain. In such cases this pain could be alleviated by
reducing power to a specific segment or segments.
The resulting treatment temperatures in the groin and body trunk were higher than
those of other sites. These regions lend themselves to better acoustic coupling and alignment
than the headlneck and axilla. Treatments of the head and neck site were less successful
than other sites. This may be attributed to several factors including: relatively higher blood
flow, limited anatomical US windows, difficulty of coupling the applicator into this region
and limited patient tolerance due to pain and position. Other investigators have also reported
less success inducing hyperthermia with US in the head and neck site (Shimm et al. 1988).
The treatment set-ups associated with the Sonotherm system are difficult and time-
consuming. The applicators are heavy and bulky. The weight and size make it difficult
to position on h e patient with the desired spatial alignment. Three different coupling bolus
techniques were tried; however, all approaches have significant limitations. Patients can
and do easily shift their relative position with respect to the applicator during treatment.
This compromisesacoustic coupling as well as the geometric correlationbetween individual
US segments and location of temperatureprobes. Difficulties in maintaining good geometric
alignment and acoustic coupling, together with the fact that pain symptoms and not
temperature often limit treatments, made the system's automatic temperature control
algorithm of marginal use in our experience.
Temperature measurements in US hyperthermia have known uncertainties (Fessenden
et al. 1984). The current ATS acquisition technology and the Sonotherm thermometry
system and software do not allow a real-time means for accurately correcting artefact
associated with temperature measurements in the presence of the US field. Therefore, no
correction has been applied to the data and analysed results. We estimate that the viscous
artefact error associated with the temperature mapping device and fixed thermocouples
would bias these data towards higher measured temperature by a few tenths of a degree
Celsius. (Absorption errors associated with vinyl-coated surface probes are much higher
and these surface data are not included in the results.) Errors associated with thermal
conduction along the metallic mapping needle were found to be similar to those found
for a non-metallic probelcatheter combination routinely used in MW hyperthermia.
Multi-element ultrasonic hyperthermia system 921
thermometry (Lyons ef al. 1985,Samulski et al. 1985). Therefore no attempt to correct
for thermal conduction smearing was made to the data.
Improvements in both hardware and software for the Sonothenn system are forthcoming.
This set of data based on 147 treatments of 31 patients will be used to document
quantitatively the impact of such improvements on our ability clinically to achieve higher
temperatures.
Acknowledgments
The authors appreciate contributions from Jeanne Forest and Dennis Lawson. This
work is supported in part by NCI/NIH grant no. 5P01 CA42745-03 and the Walther
Medical Research Institute, Inc.
References
COLEMAN, D. T., LISSI,F. L., BURGESS, S. E. P., SILVERMAN, R. H., SMITH,M. E., DRILLER,
T., ROSADO,A., ELLSWORTH, R. M., HAm, B. G., ABRAMSON, D. H., and MCCORMICK,
B., 1986, Ultrasonic hyperthermia and radiation in the management of intraocular malignant
melanoma. American Journal of Ophthalmology, 101, 635-642.
CORRY,P. M., BARLOGIE, B., TILCHEN, E. J., and ARMOUR,E. P., 1982, Ultrasound induced
hyperthermia for the treahnent of human superficial tumors.International Journal of Radiation
Oncology, Biology, Physics, 8, 1225-1229.
CORRY,P. M., JABBOURY, K., ARMOUR,E. P., and KONG,T. S., 1984, Human cancer treatment
with ultrasound. IEEE Transactions in Sonics and Ultrasonics, 31, 444-456.
CORRY,P. M., JABBOURY, K., KONG,T. S., ARMOUR,E. P., MCGRAW,F. J., and LEDUC,T.,
1988, Evaluation of equipment for hypedermia treatment of cancer. InternationalJournal
of Hyperthennia, 4, 53-74.
ENGLER,M. J., DEWHIRST,M. W., WINGET, J. M., and OLESON,J. R., 1987, Automated
temperature scanning for hyperthermia treatment monitoring. International Journal of Radiaton
FESSENDEN, P., LEE,E. R., and SAMULSKI, T. V., 1984, Direct temperature measurement. Cancer
Research (Suppl.), 44, 4799-4804.
HAND,J. W., and JAMES,J. R., 1986, Physical Techniques in Clinical Hyperthennia (Research
Studies Press, Letchworth).
HYNYNEN, K., SHIMM,D., ROEMER, R. B., ANHALT,D., and CASSADY, J. R., 1987, Temperature
distributions during clinical ultrasound hyperthermia. Proceedings of the IEEE 9th Annual
Conference of the Engineering in Medicine and Biology Society, 3, 1644-1645.
KAPP,D. S., FESSENDEN, P., SAMULSKI, T. V., BAGSHAW, M. A., Cox, R. S., LEE, E. R.,
LOHRBACH, A. W., MEYER,J. L., and FWONAS,S. D., 1988, Stanford University
institutional report. Phase I evaluation of equipment for hyperthermia treatment of cancer.
International Journal of Hyperthermia, 4, 75-1 15.
LELE,P. P., 1983, Physical and clinical studies with ultrasonic hyperthermia. Hyperthennia in
Cancer Zkrapy, edited by F. K. Storm (G. K. Hall, Boston, MA) pp. 333-367.
LELE, P. P., 1986, Rationale, technique and clinical results with scanned, focused ultrasound
(SIMFU) system. Proceedings of the IEEE 8th Annual Conference of the Engineering in
Medicine and Biology Society, 3, 1435-1440.
LYONS,B. E., SAMULSKI, T. V., and BRITT, R. H., 1985, Temperature measurements in high
thermal gradients: I. The effects of conduction. International Journal ofRadution OncoCogy,
Biology, Physics. 11, 951-962.
MARMOR, J. B., POUNDS,D., Bosnc, T. B., and HAHN,G. M., 1979, Treatment of superficial
human neoplasms by local hyperthemia induced by ultrasound. Cancer, 43, 196-205.
OLESON, J., and SAMULSKI, T. V., 1989, Improved coupling of ultrasound hyperthermia applicators
to patients. International Journal of Radiation Oncology, Biology, Physics, 16, 609-612.
SAMULSKI, T. V., LYONS,B. E., and B m , R. H., 1985, Temperature measurements in high
thermal gradients: II. Analysis of conduction effects. International Journal of Radiation
SAPOZINK, M. D., GIBBS,F. A., EGGER,M. J., and STEWART, J. R., 1986, Regional hyperthermia
in the treatment of clinically advanced deep seated pelvic malignancy. American Journal of
Clinical Oncology, 9, 162-169.
SHIMM,D. S., HYNYNEN,K. H., ANHALL,D. P., ROEMER,R. B., and CASSADY,J. R., 1988,
Scanned focused ultrasound hyperthermia. Initial clinical results. Interndonu1 Jouml of
Radiation Oncology, Biology, Physics, 15, 1203-1208.
UNDERWOOD, H. R., B m m , E. C., OWELTREE,K. B., and MAGIN, R. L., 1987, A multi-
element ultrasonic hyperthermia applicator with independent element control. Zntemafionul
J o u m l of Hyperthe&, 3, 257-267.

Samulski 1990

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Samulski 1990

Uploaded by

Copyright:

Available Formats

INT. I. HYPERTHERMIA, 1990, VOL. 6, NO.

Clinical experience with a multielement ultrasonic hyperthermia system:

02566736/90$3.00 01990 Taylor & Francis Ltd.

2. Materials and methods

The Sonotherm loo0 system is a product of Labthermics, Inc. Technical aspects of

2.3. Patient population

Patients having biopsy-proven primary, recurrent or metastatic malignancies of various

23 Adenocarcinoma 6~6x3157 2 40.90

2.4. Treatment set-ups

respectively. The artefactwas estimated by extrapolatingthe linear temperaturerise response

4.2. Temperature analysis

compared by number of patients within these site categories.

You might also like